At VLDB 2026 in Boston, Koumudi Ganepola presented the paper Computing Why-Provenance for Property Graph Queries, co-authored with Maxime Jakubowski and Katja Hose.

This work investigates why-provenance in the context of GQL (Graph Query Language), the standard query language for Labeled Property Graphs (LPGs). In this paper, the authors provide a formal definition of why-provenance for GQL queries, identifying the fine-grained components of an LPG that contribute to a query answer. They also present an efficient query-rewriting-based algorithm for computing why-provenance. The approach leverages the native query execution capabilities of existing GQL-compliant graph database systems, enabling provenance computation without requiring a separate specialized execution engine. This work establishes a foundation for further research on provenance in the property graphs and GQL.

Download the paper (PDF): https://www.vldb.org/pvldb/vol19/p3552-ganepola.pdf